An average time-to-hire of 40 days. Hiring costs in excess of $2,000 per candidate. An average turnover rate of 60-70%.
The challenges of hourly recruitment in the retail industry have been well-documented.
Despite this, many of the largest companies persist with old-school recruitment processes.
Given the break-neck pace and scale of the industry, it’s hard to diagnose and fix the problem.
Understandably, many HR leaders have been quick to layer on technology solutions that seem to make things easier; in actuality, these tech solutions have added complexity, making efficiency gains difficult and actionable insights hard to find.
Where recruitment is concerned, a HR tech stack tends to look like this: an unwieldy ATS, often coupled with a conversational AI or scheduling tool.
This stack is implemented across a decentralized system – hundreds of stores across the country – resulting in a situation where hiring managers are forced to use systems they don’t understand and don’t like.
The bottom line is this: Retail companies are overstacked, overworked, and need to adopt different solutions to old problems.
One of the biggest challenges with recruitment at major retail companies is high turnover rates. Retail staff members move fast and often, and have a high likelihood of migrating to competing businesses.
This is partially a nature-of-the-beast problem, but if we better understand what makes people tick, we can better match them to the roles at which they’re likely to succeed, and therefore keep them longer.
For example, we know that the best retail cashiers are high in extraversion. They’re energized by being around people, have good interpersonal skills, and have a lower likelihood of experiencing negative emotion while on the job.
It makes sense, then, to prioritize extraversion when matching candidates to the role of cashier. That’s a personality trait – with attendant soft skills – that will predict success for that role.
When people are matched to the job for which they are best suited, they’ll experience higher levels of purpose and satisfaction. It’s obvious why – the daily activities will invigorate rather than drain them. People who have purpose stay longer.
Therefore, if you accurately match soft skills to roles, you’ll reduce churn. Our AI Smart Chat Interviewer is really good at this: Across the board, our skill-matching power reduces non-regrettable churn by a minimum of 25%.
If you’re keen to get started measuring soft skills, download our HEXACO job interview rubric. It features more than 20 interview questions designed by our personality psychologists to assess the skills of candidates that come your way. It will even help you figure out what soft skills are best for you based on the needs and values of your organization.
Chances are, when your employees or candidates leave, they’re probably staying within the industry – and that means they’re likely going to your competitors. It’s 2023, and the stock-standard advice would be to offer higher wages and perks.
That’s not always feasible, and besides, there’s no guarantee that doing so will markedly reduce the threat of poaching and abandonment. Money is important, but it doesn’t trump purpose and belonging.
The key to better employer branding is a system for active listening. Find out what your people, be they employees or candidates, think. Ask them often. It’s important to do this at the onboarding stage, but it should continue through to the point of highest churn – the six-month mark.
Our joint report with Aptitude Research uncovered some interesting data on the importance of two-way feedback between candidates and employers.
Gathering and acting on mutual feedback:
An NPS (Net Performer Score) framework is a good place to start. How likely are you to recommend our company to a friend or colleague?
The NPS tracking question is easily configurable and embeddable into automated emails, meaning it can be set up through your ATS with little additional work.
When you begin to analyze the data, keep things simple: Dump the data into a spreadsheet, and look at your average numbers. If your score is below 0, you’ve got work to do – if it’s 0 to +30, you’re doing well. 30+ and over, well done!
(If you’re reading this, it’s probably not likely that you’ll get a 30+ score on the first go-round. That’s okay – the goal is to find out how much work you’ve got to do.)
The benefit of benchmarking NPS is that it gives your business a single, easy-to-understand proxy for employee engagement. Once you’ve got the number, you can start to make small changes and see how that affects the overall number.
We hear it all the time: Sourcing is a big problem. When we ask customers about their current processes, however, a common problem emerges: We don’t really know how many people we’re losing from our recruitment funnel, and why.
This presents a great opportunity: Often, improving an application process means removing things, rather than adding them.
Conventional wisdom tells us that the longer your application and interview process goes on, the higher your dropout rate will be. But that’s a generalized issue – it tells you nothing about how to fix the problem, beyond simply making it shorter. You need specific, localized data to diagnose and fix your leakage spots.
Data from a 2022 Aptitude Research report on key interviewing trends found that candidates tend to drop out at the following stages, in the following proportions:
Let’s say that you had 100 visitors to your careers (or job ad) page, and 20 of them completed the first-step application form on that page. You’ve lost 80% of your possible pool right there. Not great, but at least you know – now you can examine that page to uncover possible issues preventing conversion.
Is the page too long? Does it have too much text? Is the ‘apply’ button clearly shown? Is the form too long, requiring too much information to fill out? Are your perks/EVP attributes clearly displayed?
Having been a CHRO of a listed company in my last role, I can empathise with the confusion and exhaustion that comes from navigating the myriad HR tech products flooding the market whilst trying to manage many ongoing HR change initiatives.
Last year, as CEO of an HR tech start-up I did what most do in that role — I spent a whole lot of time talking to customers, CHROs, heads of talent, recruiters and business owners, listening to their challenges to build a product that works for them. There are a few themes I picked up on through these conversations.
‘What’s the right tech stack for my team and our company?’ and ‘how do I integrate all these technologies?’ are questions every CHRO of any sizeable company is grappling with. And the answer is more complicated than committing to a new HRIS.
Whilst I am not a tech expert, I spend many hours a week thinking about one critical part of the HR function that is ripe for technology innovation — recruitment. In that vein, I am sharing some things I have learnt which I hope will be useful to your investments in your tech stack in 2019.
There are HR tech products that give you insights on engagement hot spots, employee sentiment, and screen applicants for roles by scraping and analysing people’s personal profiles or communications. If you believe (as I do) that transparency enhances trust, especially when it comes to anything coming out of HR, these tech products could undermine organisational trust and maybe even your employer brand. Look beneath the hood of a tech product to validate how it works. AI and the concern of algorithmic bias is one every CHRO needs to be ready to talk about. Understand the source data and how it will be used in the solution. For candidate selection, any front end testing needs to not only be valid but feel valid to the user. That’s why we use relatable and valid questions to assess candidates in building our predictive models. No CVs, no video and no games.
Any extra discretionary effort by employees is going to be heavily influenced by how much trust your people have in you. Better to invest in tech solutions that allow for more transparency around how decisions are being made, that use reliable, objective and valid data.
Think of the people analytics generated by HR today — turnover reports, engagements stats, culture diagnostics, exit survey analysis, 9 box talent management. All of it is backward-looking reporting on the past performance of talent. Much of it also subject to the vagaries of human analysis, therefore biased insights. How many of your organisations use data to validate the placement of people on the ‘potential axis’ of a 9 box? Or use NLP to extrapolate the key themes from engagement surveys and exit survey verbatim?
A bigger challenge for all of this backwards analytics is connecting the dots — how does a culture survey actually move you towards and predict a different culture? My colleague who spent his early years building up the data science team for a leading engagement survey platform and led the benchmarking analysis for their clients observed that year after year the same companies were in the top and the bottom quartile of engagement.
Changing culture is hard unless you change the people — the people you hire and the people you promote.
The best investment you can make to change the culture and help the organisation move towards forward-looking predictive analytics is to start to capture data from the outset — from your applicant pool, through to the people you hire.
Having a data DNA profile of your applicant and hired pool means you can better target your employer branding, you can identify with high accuracy the profile of the stronger performers, the people who are high flight risk in the early months, the talent that moves fastest to productivity. Knowing these profiles means you can seamlessly feedback into your recruitment a better hiring profile.
This is the power of predictive analytics over psychometric testing which has no feedback loop back to the business on whether the person with the high OPQ test was any good in the role.
‘Garbage in garbage out. This is usually a reference to a data quality issue.
Data can take many forms- it’s not always hard numbers (more on that later), it can be data that is structured and regulated by you vs data that is unstructured and not regulated by you, such as CV’s. The former is always better — closer to the objective source of truth, usually owned by you, and less prone to gaming.
CVs are a poor man’s data substitute and rarely indicative of anything. A CV is a highly gameable type of data and relying on CV data to select talent exacerbates the risk of bias, as was experienced by Amazon when they built their hiring models around a 10-year database of CVs (mostly male).
I won’t spend time on the risks of bias in CV screening as enough has been written about that, other than to share this from a blog post which quotes academic research that ‘both men and women think men are more competent and hirable than women, even when they have identical qualifications ‘, and that ‘resumes with white-sounding names received 50% more calls for interviews than identical resumes with ethnic-sounding names’. https://www.lever.co/blog/where-unconscious-bias-creeps-into-the-recruitment-process.
Removing bias in the screening process is no longer about social justice, now it’s about commercial outcomes — McKinsey has documented each year since 2014 that companies with top quartile diversity experience outsized profitability growth https://www.mckinsey.com/business-functions/organization/our-insights/delivering-through-diversity
There are a plethora of surveys that make the point that HR functions are starting to invest in the power of people analytics.
Making data more visual has been a big driver behind the success of engagement analytics companies such as a Culture Amp, Glint and Peakon, transforming ugly engagement decks and the traditional circumplexes into insights-driven real-time dashboards. Visualisation offered by tools like Tableau is table stakes these days for HR.
Data doesn’t always look like data in a traditional sense. Take textual search data, human behavioural tracking data for example. Google has been making money off that data strategy for years and there are now books written about how google search terms are the most accurate mirror to our true beliefs and values (Read Everybody Lies for a fascinating insight into the power of text).
Tracking human behaviour has been mainstream in marketing teams for years, but has been slower to be leveraged in HR. In consumer marketing, no one cares why a person is more likely to buy an item, they are only interested in optimising for the outcome. There has been some interesting research applying consumer behaviour analysis to HR with fascinating insights, for example, that your choice of browser in completing an online assessment is a strong predictor of your performance in the role.
In consulting there is an often-used accusation of consultants ‘boiling the ocean’, which usually refers to those 100-page decks with chart after chart, visualising every data point possible as if the sheer weight of the deck is somehow testament to its accuracy.
Most junior consultants aspire to write the ‘killer slide’, the elusive one slide that crystallises the strategy in one data visual that will transform the company’s trajectory.
As HR teams start to produce more output on people analytics, there is a risk of ‘boiling the ocean’ on people analytics — quarterly engagement surveys, monthly churn data, diversity reporting. Figuring out the ‘so what’ of the data and using those insights to move the needle on business metrics that matter is harder, but also necessary. For HR integrating non-HR owned data is also important to get a fuller picture, especially for sales led businesses. For example, if sales drop off at the 2-year mark, what can HR do about that? What HR processes change as a result of seeing high correlations between sales trajectory in the first 6 weeks and tenure greater than 6 months.
HR’s role is very much one of building bridges across the organisation — taking a helicopter view of talent, ensuring that the needs of the business will be met in 3 years, 5 years by the people in the business, in enabling communication and collaboration channels across teams and geographies.
Building a single source of truth about their employee base often justifies HR’s biggest tech investment in helping achieve those objectives — the so called ‘one size fits all’ HR system. Yet it’s a big step to assume that even with the HRIS in place that HR has all the data it needs to do its job. Every function is making similar investments — sales & marketing into CRMs, operations teams into rostering systems, LTI and OHS data that might sit in the BU or a separate OHS team.
Last century, HRs accountability might have ended when they filled a role. Today, HR is accountable for ‘talent optimisation’ and that means ensuring people’s success through their career with the organisation, and often even beyond. Knowing how that talent is performing on the job– roster adherence, injury patterns, call centre metrics, sales performance — are integral to optimising that talent pool.
Capitalise on these various streams of data!
I encourage HR leaders to be expansive about what is performance data, especially objective performance data, and being relentless in sourcing that data from their non-HR colleagues internally.
Data generated within HR can help drive broader organisation decisions. B2C companies with large volumes of sales and marketing applicants can leverage the power of those volumes for the benefit of the rest of the business.
Big brand companies can receive half a million-plus applications in one year, often engaging meaningfully with just a fraction. Technology allows you to test and engage meaningfully with every one of those applicants. Instead of thinking of that pool only as a candidate pool relevant to recruitment, for a B2C business, that pool is most likely also your consumer base and a rich source of data for your business.
Customer acquisition cost (CAC) for product and services like travel, retail, software, financial products range from $7 to $400, with companies committing substantial advertising budgets to reach that kind of audience, yet over in recruitment, they are engaging with them for free, at a point where the candidate/consumer is at their most willing and motivated to engage with you.
Imagine what consumer data you could capture from that applicant pool for the benefit of the business?
Transparency and authenticity, forward-looking predictive data, business impact first, think creatively and broadly, and HR as a data generator. These are 7 themes that can transform your organisation in by leveraging the data hidden within HR through the efficient use of technology.
Sapia (Previously PredictiveHire) has won Best Innovation in Algorithmic Bias Mitigation for Fair AI in Recruitment (FAIR™). CogX is the world’s largest celebration of Ai and emerging technology. This is the second time we have been honoured at these awards shortlisted last year for Best Conversational Ai Solution for HR.
The big question of this year’s CogX awards “How do we get the next 10 years right?” couldn’t be more important to us. We want to give everyone a fair go when it comes to getting a job. It has the potential to have so much impact in addressing bias and structural inequality in the world. We know because we’ve seen it. We’ve helped some of the world’s most trusted consumer brands achieve their DE&I metrics by giving all applicants a fair go.
We know Ai tools can seem confronting in concept. Ai has an incredible ability to provide reliability and comfort that outcomes are fair, but if not implemented and used correctly, it can compound bias. Because Ai technology is outstripping regulation in most countries, poor algorithmic hygiene can easily creep in.
We believe that our platform is the world’s most inclusive talent solution and we’re proud to be leading the way on the ethical use of Ai for hiring. We want to empower leaders considering Ai for hiring by giving them the right questions to ask vendors, and we want to get developers talking about bias mitigation best practices.
This year we made FAIR™, our framework for the ethical use of Ai in hiring, available on the public domain. Our framework presents a set of measures and guidelines to implement and maintain fairness in Ai-based candidate selection tools. It is a data-driven approach to fairness. It was a bold move, but CogX liked it.
FAIR™ was created by our team of incredibly dedicated data scientists, led by the incredibly humble Buddhi Jayatilleke our Chief Data Scientist. Our team have tested and re-tested and experimented and re-experimented to find a new formula for assessing talent–one that is 100% inclusive and bias-free.
The winners of the CogX Awards were announced during CogX 2021 in London, June 14-16, 2021.
To find out how to interpret bias in recruitment, we also have a great eBook on inclusive hiring.
It’s the start of the football season in Australia, and I’ve been thinking about how damn hard it must be to coach. Every week you need to revise your game plan, pick the best team to get the match-up right and galvanise your players behind your decisions. You’re constantly adjusting your approach depending on who you are up against, picking different players to counter your opponent’s strengths.
I have heard it said by people who are more football savvy than me that the best coach is the one who can rid him/herself of biases at the decision point of picking the right players for the match.
The well-known Nobel Laureate behavioural economist, Dr Daniel Kahneman, made the same discovery when he first started to work in the space of ‘human decisioning’.
As a young psychologist in the 1950’s, he was tasked with figuring out which recruits to the army should be allocated to which units, infantry, air force etc.
All the generals of those units strongly asserted that there was a different type for each unit, and they wanted to make allocations that reflected those differences. What Kahneman found over time was that there was no difference between the best soldiers in each part of the army.
What he observed was the intrusion of the interviewer’s own intuition when interviewing for these roles. Expert judgments were less reliable than they thought. The right algorithms might have solved this.
Accuracy is not correlated with experience. We rely on heuristics, rules of thumb, mental models that rely on similarity with our own past experience.
The more likely the individual in front of you looks like your mental model of the person you hired last time that “was a great (insert role)”, the more likely you are to see them as a great (insert role).
Confirmation bias is something we see every day in the stock market. A stock price increase does not mean a company is successful. These buying decisions are emotional, not rational. People decisions suffer from the same human frailty.
That’s why Nobel Laureates like Dr Daniel Kahneman and many smart HR leaders are recognising that the only way to interrupt our own biases is with the right technology.